VGG Image Annotator

VGG Image Annotator is a free browser-based suite for manually labeling images, audio and video. It is best suited to researchers and teams who want flexible annotation and export tools without installation. Its focus on manual work makes it a poor fit for anyone seeking model-assisted labeling.

Overview

VIA runs from a standalone HTML file in a browser, including offline, so annotators can work without installing software or setting up a server. That is useful for self-contained projects and environments where an internet connection is not part of the workflow. The trade-off is that VIA provides manual annotation rather than model-assisted labeling.

VIA3 handles images, audio and video, and supports shared projects. VIA2 remains maintained, but is limited to images and does not support sharing. Teams doing collaborative work should choose VIA3 and connect it to a Version File Store (VFS): VGG maintains a public free instance, or an organization can run its own.

Key features

VIA supports annotation shapes including rectangles, circles, ellipses, points, polygons and polylines. That range suits projects requiring marked regions or points, but the annotation itself is manual. CSV and JSON import and export, conversion scripts for COCO and Web Annotation, and a Python package for converting VIA projects to COCO give teams several ways to move annotations into other workflows.

The software is built with HTML, CSS and JavaScript, using standard modern browser features and no external libraries. The makers report testing it in Firefox, Chrome and Safari, and say users have also found it works in other browsers. Downloadable ZIP archives and standalone HTML copies are offered for both VIA2 and VIA3. An in-app guide, FAQ and email contacts provide help. VIA is released under the permissive two-clause BSD license for academic and commercial use; the project describes offline browser operation but makes no security certification or compliance claims.

Pricing

VGG Image Annotator costs nothing. It is released under the two-clause BSD license, with no paid tiers or usage limits stated. There is no free trial because the software is free. That makes it suitable for academic and commercial projects without a subscription decision, though teams needing automated labeling will need another tool.

Platforms

VIA is listed for Linux, macOS, Windows and the web. Because it runs as a browser application distributed in a single HTML file, its practical fit depends on a supported browser; the makers report testing Firefox, Chrome and Safari.

Who it's for

VIA3 is the relevant choice for audio, video or shared projects. VIA2 remains an option for image-only work, but its lack of sharing rules it out for collaborative workflows. Neither version suits teams that want model-assisted labeling.

Pros and cons

• Offline use without installation: A standalone HTML application can run in a browser without setup, which suits self-contained workflows.
• Broad annotation and export options: Multiple shapes, CSV and JSON, plus COCO and Web Annotation conversion routes, support varied manual labeling projects.
• Free and permissively licensed: The two-clause BSD license allows academic and commercial use, with no paid tiers or usage limits stated.
• No model-assisted labeling: Teams seeking AI-supported annotation will need a different product.
• Collaboration depends on VIA3 and VFS: VIA2 does not support shared projects, and VIA3 sharing uses a public or self-run Version File Store.

Alternatives

For a broader image annotation software comparison, consider the AI image annotation tools list if model assistance is a priority. The AI data labeling tools list is another route for readers seeking AI-oriented labeling. For video-oriented discovery, browse AI video search tools; for a wider labeling category, see data labeling software.

FrameQuery is worth considering when video search and sharing across local storage, Drive or NAS matter more than annotation; its free plan has no monthly indexing hours, and face, voice and AI-tag search require indexed video. Choose DISKO for screenwriting and shot-list tools with review links; its free cloud plan includes 2 GB. Cinestar is a free open-source core for Linux and macOS, with local AI services requiring Ollama.

Moment Search suits macOS and Windows users who want on-device visual and speech search, with a free starter plan capped at one footage source and 25 indexed videos. QMedia is a free, MIT-licensed option for API, self-hosted or web use. Scout Suite is a free open-source multi-cloud security auditing tool, rather than an annotation product.

Twelve Labs offers video indexing through API and web, with its free plan capped at 600 minutes, 90-day index access, 10 hours per index, 100 videos per index and five concurrent indexing tasks. ViSearch is a paid video-processing option with a free allowance of up to 20 minutes and a Basic rate of $0.09, paid once, billed $0.09/min.

Verdict

Choose VIA if your priority is free, offline manual annotation with several useful export paths, especially for research or a workflow that benefits from VIA3 sharing. Look elsewhere if you need model-assisted labeling or a product built around video search rather than annotation.